LECTURE 2: BASICS OF EXPERIMENTAL
DESIGN
Experimental design sounds a lot more technical
and trickier than it actually is.
You use basics of experimental design every day,
although you might not recognize it.
BI/EN 481: RESEARCH METHODS AND
SEMINAR
LECTURE 2: BASICS OF
EXPERIMENTAL DESIGN
For example, if you want to learn
something like if a handsome or
pretty classmate likes you, you
might try to come up with a good
“experiment” to find this out.
You might decide on an
“experiment” where you ask the
student if they would like to
study with you or, go to town
with you.
BI/EN481: RESEARCH METHODS AND
SEMINAR
We design experiments all the time--
anytime we want to find out some
sort of new knowledge.
Is the water hot enough for tea? A bad
experimental design might be to dip
your finger deeply in the water, Ouch,
yes, it is ready, but now I’m burned.".
A better design might be to look at the
water and see the bubbles on the
bottom of the pot: Yes, its beginning to
boil and I’m not burned.
BI/EN481: RESEARCH METHODS AND
SEMINAR
Experiments are not just
things boring scientists in
white lab jackets do inside
laboratories.
With a little knowledge of the
situation–
the physics of heat transfer or
what your attractive classmate
likes to do-- you could design
the best "experiment”
BI/EN481: RESEARCH METHODS AND
SEMINAR
The same is true with biology.
With some fundamental
knowledge about the things you
want to learn about, (your basic
biology) and some design tools
(this course), you will be prepared
to go out and discover new things.
This is the time (final year) to put
to good use your basic knowledge
of biology!
BI/EN481: RESEARCH METHODS AND
SEMINAR
Make the most of this course to equip yourself with the
basic research skills!
You might not have the chance again to do a research
design course until you find yourself with a research
project as a career as you graduate.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
2.1 Hypothesis testing
Like "experiential design,"
hypothesis testing sounds a bit
technical and not like something
you do very often' but indeed, we
do it all the time-- anytime we
formulate an idea about
something.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
In the previous example we
might have formed a
hypothesis that our
attractive colleague does
or does not like us.
Hypothesis is like a guess
for which you needed a
good research design to
make a good conclusion.
BI /EN481: RESEARCH METHODS AND
SEMINAR
You decide to test this hypothesis by seeing
if s/he want to spend some time with you.
This is a reasonable design to test your
hypothesis. But it can still be tricky.
For example, you invited him/her to go
fishing on your uncle’s boat, you might
have been turned down because s/he gets
seasick. You might mistakenly interpret this
as dislike.
It was a bad experimental design. You
tested the hypothesis that the person
likes to fish, not whether they wanted to
be with you.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
Hypothesis testing means that we
clearly and unambiguously create a
question or a series of questions that
we have discrete answers.
We then conduct an experiment to get
an answer and then we can draw
conclusions about our hypothesis.
The fishing example was not a good
design because we confused several
hypotheses (attraction to you and
attraction to being on the boat).
BI /EN 481: RESEARCH METHODS AND
SEMINAR
You might have solved this with a
design where you then asked your
attractive classmate if s/he would rather
go get some food with you, go to the
library with you or to a shop with you,
etc.
After being told “no” to all these things,
you might begin to feel you’ve
answered the hypothesis and can
conclude that the person does not really
want to spend anytime with you.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
At this point we usually revise our opinion
and decide that the other student is not
attractive after all!
Good experimental design does not
mean that we always get the answer we
desire!
BI /EN 481: RESEARCH METHODS AND
SEMINAR
Experiments are usually designed to
test hypothesis.
What is a hypothesis?
BI /EN 481: RESEARCH METHODS AND
SEMINAR
What is a hypothesis?
A hypothesis is anything we might
think be happening. It is a proposed
idea intended to explain certain facts
or observations
BI /EN 481: RESEARCH METHODS AND
SEMINAR
We can examine observational data to
test hypothesis, in other words, to see if
the data support our hypothesis or not.
For example, we could make the
hypothesis that there are more Eurasian
Tree Sparrow (Passer montanus) on the
south side of UoG campus than on the
north side of the campus.
We could test this hypothesis by
comparing population counts from
both sides.
This is an example of hypothesis testing
using observational data.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
Or, we could make the hypothesis
that lowland forests have more
rattan (Calamus spp.) than
highland forest.
We could test this hypothesis by
comparing vegetation plots from
both altitudes.
This is another example of
hypothesis testing using
observational data.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
Even survey work is designed to
test hypothesis although this
might not be apparent.
Basically, any survey assumes a
hypothesis that the ……..(frogs,
trees, birds, bats, etc) at one site
are different or not different from
the …………..(frogs, trees, birds,
bats, etc) at another site.
Biodiversity surveys are basically
inventory of species diversity
BI/EN481: RESEARCH METHODS AND SEMINAR
BI /EN 481: RESEARCH METHODS AND
SEMINAR
So although survey work
seems purely descriptive,
it is intended for use in
hypothesis testing.
Consequently, we need to
collect our survey data in
a consistent, scientifically
rigorous fashion just like
experimental data.
If we do not, we will not
be able to statistically test
hypotheses
BI /EN 481: RESEARCH METHODS AND
SEMINAR
Project design can encompass much more than just basic
surveys to see if you have more of one thing at one site than
another.
We can study animal behaviour, plant growth rates, physiology
etc. Anything at all in the biological sciences, either in the field
or in the lab. The skills of good design and analysis are
applicable everywhere.
But before we can begin to design any sort of project, we must
make sure we have a good well- defined hypothesis. This is
the essential trick to doing good science.
BI /EN481: RESEARCH METHODS AND
SEMINAR
A good hypothesis is characterized by three qualities:
1. Unambiguous: It is generally difficult to simultaneously
test multiple hypothesis. So, you want to define how you
word your hypothesis so that it is as close as possible to one
single hypothesis and is not confused with another hypothesis.
For example, you want to learn if someone wants to spend
time with you, and not whether they like fishing.
BI/EN 481: RESEARCH METHODS AND
SEMINAR
A good hypothesis is characterized by three qualities:
2. Explicit and clear: Once you have focused on a single
hypothesis, you need to define it precisely; sloppy language
will lead to a sloppy design. Even if in your mind, you know
what you want to test, you must state it clearly. For example,
you wanted to know if the water was hot enough to brew a tea,
not just whether it was “hot” a term that is relative.
BI /EN 481: RESEARCH METHODS AND
SEMINAR
A good hypothesis is characterized by three qualities:
3. Testable: It is not science to make hypothesis that are not
testable. If the fishing is poor, you might blame the spirits of
the ocean but unless you have some way of testing to see if the
spirits are responsible, it is a poor hypothesis an is beyond the
realm of science.
from the above infromation given, how can they set some test questions at university level.
At a university level, test questions could be set based on the concepts of experimental design and hypothesis testing. Here are some examples of test questions that could be asked:
1. Define experimental design and provide an example of how it can be applied in a real-life situation.
2. Explain the importance of having a well-defined hypothesis in scientific research. Give examples of characteristics of a good hypothesis.
3. Describe the differences between a bad experimental design and a good experimental design using specific examples.
4. Discuss the role of hypothesis testing in scientific research. How does hypothesis testing help researchers draw conclusions from their experiments?
5. Analyze the potential pitfalls in experimental design and how they can affect the validity of research results.
6. Compare and contrast observational data and experimental data. How are they used in hypothesis testing?
7. Create a hypothetical research study and outline the steps you would take to design an experiment to test a specific hypothesis.
8. Evaluate the importance of collecting data in a consistent and rigorous fashion for hypothesis testing. How does data collection affect the reliability of research findings?
9. Discuss the applicability of experimental design and hypothesis testing in various fields of biology, including animal behavior, plant growth rates, and physiology.
10. Formulate a hypothesis related to a specific biological phenomenon and propose an experimental design to test this hypothesis. Provide a detailed explanation of your experimental approach.